论文部分内容阅读
主成分变换对于区分多光谱遥感数据中岩性信息有一定效果。利用主成分变换的２ 个因子散点图分类提取多光谱遥感数据中较弱岩性信息时，发现这些弱信息并不是出现在散点图重合部分的重心位置，说明多光谱原始遥感数据之间除了存在着起支配作用的主特征值外，还存在反应较弱信息的特征因子。而主成分变换只表示了特征空间系列变换的一种结果，为了提取和显示多因子变化空间中出现的较弱特征信息，开发了多因子逐步正交变换算法软件，在峪耳崖金矿带应用时取得了较好的效果。研究中还采取了掩膜技术排除水体和河道的干扰信息。该方法已用在华北地台北缘其它地区的找矿。
Principal component transformation has certain effect on distinguishing lithology information in multispectral remote sensing data. When we extract the weaker lithologic information from multispectral remote sensing data by using two factor scatterplot of principal component transform, we find that these weak information do not appear in the center of gravity of overlapped part of scattergram, indicating that between multispectral remote sensing data In addition to the dominant eigenvalues that are dominant, there are also characteristic factors that respond to weaker information. However, the principal component transformation only represents a result of series transformation of the feature space. In order to extract and display the weak feature information appearing in the multi-factor change space, a multi-factor step-by-step orthogonal transformation algorithm software was developed. Application has achieved good results. The study also adopted a mask technology to exclude interference information of water and river. This method has been used for prospecting in other parts of the northern margin of the North China Platform.